Search

Background Individuals with bipolar disorder are commonly correctly diagnosed a decade after symptom onset. Machine learning techniques may aid in early recognition and reduce the disease burden. As both individuals at risk and those with a manifest disease display structural brain markers, structural magnetic resonance imaging may provide relevant classification features. Methods Following a pre-registered protocol, we trained linear support vector machine (SVM) to classify individuals according to their estimated risk for bipolar disorder using regional cortical thickness of help-seeking individuals from seven study sites (N = 276). We estimated the risk using three state-of-the-art assessment instruments (BPSS-P, BARS, EPIbipolar). Results For BPSS-P, SVM achieved a fair performance of Cohen's κ of 0.235 (95% CI 0.11–0.361) and a balanced accuracy of 63.1% (95% CI 55.9–70.3) in the 10-fold cross-validation. In the leave-one-site-out cross-validation, the model performed with a Cohen's κ of 0.128 (95% CI −0.069 to 0.325) and a balanced accuracy of 56.2% (95% CI 44.6–67.8). BARS and EPIbipolar could not be predicted. In post hoc analyses, regional surface area, subcortical volumes as well as hyperparameter optimization did not improve the performance. Conclusions Individuals at risk for bipolar disorder, as assessed by BPSS-P, display brain structural alterations that can be detected using machine learning. The achieved performance is comparable to previous studies which attempted to classify patients with manifest disease and healthy controls. Unlike previous studies of bipolar risk, our multicenter design permitted a leave-one-site-out cross-validation. Whole-brain cortical thickness seems to be superior to other structural brain features.

There is growing evidence of abnormalities of high-frequency oscillations in the gamma range of the electroencephalography in schizophrenia. The generation of neural activity in the gammaband was shown to be critically related to a glutamatergic and GABAergic microcircuit which is also known to be involved in the pathophysiology of schizophrenia. One example of such gamma oscillations is the early auditory evoked gamma band response (aeGBR). We aimed to investigate whether there are altered aeGBR and activity of its sources in the anterior cingulate cortex and/or the auditory cortex (identified as sources of the GBR previously) in schizophrenic patients and in first-degree relatives of schizophrenia patients. We investigated the early aeGBR and its sources (LORETA source localisation) in 90 medicated patients with schizophrenia and in 17 unaffected first-degree relatives of patients with schizophrenia using an auditory reaction task (comparison with age-, gender- and educational-level-matched control groups). Evoked power and phase locking of the aeGBR was reduced in schizophrenia patients and healthy first-degree relatives of patients with schizophrenia. This effect was due to a reduced activity in the auditory cortex and the anterior cingulate cortex. The findings are in line with the hypothesis of a disturbed GABAergic interneural modulation of pyramidal cells in schizophrenia and findings of different schizophrenia risk genes associated with transmission at glutamatergic and GABAergic synapses. The results regarding the first-degree relatives of patients with schizophrenia point to the applicability of this marker as a heritable intermediate phenotype for schizophrenia.

Executive functions comprise various cognitive abilities including the inhibition of prepotent responses and voluntary decisions. Several studies showed medial-frontal activations in tasks with the free selection of responses. The inhibition of prepotent response tendencies seems to be associated with medial frontal as well as lateral frontal BOLD responses. The aim of this simultaneous EEG and fMRI study was to discriminate the neural correlates of behavioural control processes in ADHD.8 adults with ADHD and 8 matched healthy subjects performed a go/nogo task comprising three different conditions: during the go condition, subjects were instructed to press a response button as fast as possible; during the nogo condition, this response was to be inhibited. In the voluntary selection task participants were allowed to freely decide, whether to press the response button or not.The fMRI protocol used a gradient-echo EPI pulse sequence. Further analyses were done with using the BrainVoyager software package (Goebel, Maastricht). EEG signals were simultaneously recorded (Brain Products, Munich).Electrophysiologically, the nogo task and voluntary decision task led to a negative decline especially in fronto-central brain regions (N2) in both groups. Regarding the functional MRI data we found inhibition-associated BOLD responses especially medial-frontal in the pre-SMA and activations in the medial part of BA 8 for the voluntary selection. ADHD patients showed a reduced contribution of frontal brain regions during free responses compared to controls.The results may indicate that selection processes are related with dysfunctions predominantly in frontal brain regions in ADHD patients.

Studies of schizophrenia with functional MRI showed hyper- and hypoactivations in various brain regions including the prefrontal cortex. Functional abnormalities have also been reported in first-degree relatives of schizophrenic patients. The aim of this study was to examine working memory related brain functions in healthy subjects, schizophrenic patients and unaffected relatives and to determine the influence of psychopathology on these processes.A parametric n-back working memory task and functional MRI were used to examine 61 schizophrenic patients on antipsychotic medication, 11 nonpsychotic relatives of schizophrenic patients and a comparison group of 61 healthy subjects. The task difficulty was incrementally increased using a parametric task (0-back, 1-back, 2-back, and 3-back) to examine the relationship between working memory load, performance, and brain activity.The results indicated that during the attention task (0-back) behavioral responses of patients and healthy subjects hardly differed but BOLD responses were considerably enhanced in schizophrenic patients. With increasing task difficulty differences between groups in BOLD responses diminished whereas behavioral deficits of patients increased. The examination of attention-independent working memory-functions (2- vs. 0-back) produced hypoactivations in patients, especially in frontal, temporal and subcortical brain regions. Behavioral performance and neural responses of unaffected relatives of schizophrenic patients were intermediate between schizophrenic patients and controls indicating slight brain dysfunctions. In addition, compensatory strategies were demonstrated.These findings suggest that the genetic risk for schizophrenia is accompanied by neural inefficiency which is associated with cognitive deficits, especially in difficult tasks.

ObjectivesThere is evidence that the antioxidant enzyme glyoxalase-1 (Glx-1) may play a role in anxiety-related behaviour. However, discordant findings between Glx-1 expression and anxiety-related behaviour have been observed in animal models. Because no data are available on the relation between Glx-1 mRNA expression and human anxiety so far, we investigated the expression of Glx-1 mRNA in peripheral red blood cells in relation to cholecystokinin-tetrapeptide (CCK-4) induced anxiety in healthy subjects.MethodsTwenty-three healthy subjects underwent challenge with CCK-4. Glx-1 mRNA expression was assessed by quantitative real-time polymerase chain reaction prior to CCK-4 injection. Baseline anxiety was assessed with the State-Trait-Anxiety-Inventory (STAI) and panic response was measured with the Panic Symptom Scale (PSS).ResultsCCK-4 elicited a marked anxiety response accompanied by a significant increase in heart rate. Glx-1 mRNA expression correlated significantly with severity of CCK-4 induced anxiety. In contrast, Glx-1 activity did not correlate with state or trait anxiety.ConclusionsThe positive correlation between Glx-1 mRNA expression and CCK-4 induced panic severity suggests that Glx-1 is involved into the acute anxiety response to CCK-4. Our preliminary findings support preclinical findings and point towards a role of Glx-1 for human anxiety.

Voluntary selection between response alternatives belong to cognitive abilities controlling and regulating goal-directed behaviour. Voluntary selection processes are associated with increased neural activity, especially in medial and lateral frontal brain regions as well as the inferior parietal gyrus. However, the precise function of each brain region as well as the spatiotemporal characteristic of the brain regions involved is not yet clear. The aim of the present study was to disentangle distinct aspects of voluntary selection and their underlying neural processes. Hence, event-related potentials (ERPs) and functional MRI data were acquired simultaneously. Brain regions modulated by the task-induced amplitude variation of ERPs (N2, P3) were indentified. The results showed N2-related hemodynamic responses, especially in medial and lateral frontal brain regions. Among other things, medial frontal brain regions are related to conflict monitoring, control of voluntary action and decision making. By contrast, the P3-amplitude proved to be predominantly related to increased BOLD responses in the temporo-parietal junction [TPJ] and lateral frontal brain regions. These brain regions are thought to play a decisive role in an attentional network involved in detecting auditory and visual stimuli.Overall, the results of the study indicated a whole network of brain regions to be associated with voluntary selection processes. In addition, at least some frontal brain regions seemed to be involved at an earlier stage than temporo-parietal regions, probably indicating a top-down process.